A major problem in optimally coordinating combination drug therapy is the inability to quantify and minimize the highly variable relationships between dosage of the various drugs, patient adherence, serum concentrations, drug -- drug interactions, shared therapeutic and toxic effects of the combination regimens given, and patient outcomes. Combination therapy is now the norm in many clinical settings. Our cross-disciplinary laboratory has developed parametric and especially nonparametric (NP) population modeling software to capture these relationships with statistical consistency and precision. We have also developed the new "multiple model" (MM) method of dosage design to hit desired therapeutic target goals with maximum precision (minimum weighted squared error), for models of single drugs having analytic solutions to their differential equations. We have now begun clinical testing in pilot collaborative projects, to make NP population models, and to achieve target goals with maximum precision. We also have NP software to make models of the larger, nonlinear and complex interacting systems of combination therapy with multiple drugs, and their shared combination therapeutic and toxic effects. Aim 1: We will implement all the above in a new Windows interface. Aim 2: We are developing MM dosage design for the combination drug regimens. Preliminary results are most encouraging. We will develop integrated software to ensure maximally precise coordinated combination drug therapy for patients with HIV, cancer, transplants, heart failure, TB, epilepsy, those requiring combination antibiotic and antifungal therapy, and even, perhaps, diabetes mellitus. Failure to consider drugs in combination means that while each drug can be individualized, the interactions are never considered, each drug appears variable and capricious, as the changing doses of the other drugs, and their effects, are not considered, and dosage adjustment is always behind the events. Our exciting new tool should now optimize the individualization and coordination of combination drug therapy for patients, with essentially optimal Bayesian MM feedback and dosage adjustment. Subsequent feedback should tend to be more confirmatory, and dosage adjustments should be fewer and smaller. 1 can also monitor effects such as Hb, WBC, platelets, viral load, CD-4, other responses, and then make adjustments of dosage to hit all selected therapeutic targets most precisely, including tolerable degrees of toxicity (Hb=10, WBC=1200, plts=100,000 for example). All this is highly feasible and most urgently needed clinically. Aim 3: This work will be studied and evaluated in several collaborating pilot clinical projects, 1 on-site, others off-site. This work should greatly improve our understanding and control of combination and interacting drug relationships, and the quality and precision of combination drug therapy for patients who must receive potentially toxic drugs. [unreadable] [unreadable] [unreadable]